You can also install Ravada using Docker.
It depends on the number and type of virtual machines. For common scenarios are server memory, storage and network bandwidth the most critical requirements.
RAM is the main issue. Multiply the number of concurrent workstations by the amount of memory each one requires and that is the total RAM the server must have. However, recent virtualization improvements allow you to overcommit the memory.
The faster the disks, the better. Ravada uses incremental files for the disks images, so clones won’t require many space.
Ravada can be installed from package or using docker.
Follow this guide if you are only upgrading Ravada from a previous version already installed.
The client must have a spice viewer such as virt-viewer. There is a package for linux and it can also be downloaded for windows.